devpts: handle bind-mounts

Bug #1755857 reported by Christian Brauner on 2018-03-14
This bug affects 2 people
Affects Status Importance Assigned to Milestone
linux (Ubuntu)

Bug Description

Hey everyone,

I sent a few patches to fix a regression caused by wrong behavior in the devpts filesystem when used with the TIOCGPTPEER ioctl() we implemented a while back. The context and reproducer can be taken from my following commit message. The required patches are in Greg's tree and are:


devpts: resolve devpts bind-mounts
Most libcs will still look at /dev/ptmx when opening the master fd of a pty
device. When /dev/ptmx is a bind-mount of /dev/pts/ptmx and the TIOCGPTPEER
ioctl() is used to safely retrieve a file descriptor for the slave side of
the pty based on the master fd, the /proc/self/fd/{0,1,2} symlinks will
point to /. A very simply reproducer for this issue presupposing a libc
that uses TIOCGPTPEER in its openpty() implementation is:

unshare --mount
mount --bind /dev/pts/ptmx /dev/ptmx
chmod 666 /dev/ptmx
ls -al /proc/self/fd/0

Having bind-mounts of /dev/pts/ptmx to /dev/ptmx not working correctly is a
regression. In addition, it is also a fairly common scenario in containers
employing user namespaces.

The reason for the current failure is that the kernel tries to verify the
useability of the devpts filesystem without resolving the /dev/ptmx
bind-mount first. This will lead it to detect that the dentry is escaping
its bind-mount. The reason is that while the devpts filesystem mounted at
/dev/pts has the devtmpfs mounted at /dev as its parent mount:

21 -- -- / /dev
-- 21 -- / /dev/pts

devtmpfs and devpts are on different devices

-- -- 0:6 / /dev
-- -- 0:20 / /dev/pts

This has the consequence that the pathname of the parent directory of the
devpts filesystem mount at /dev/pts is /. So if /dev/ptmx is a bind-mount
of /dev/pts/ptmx then the /dev/ptmx bind-mount and the devpts mount at
/dev/pts will end up being located on the same device which is recorded in
the superblock of their vfsmount. This means the parent directory of the
/dev/ptmx bind-mount will be /ptmx:

-- -- ---- /ptmx /dev/ptmx

Without the bind-mount resolution patch the kernel will now perform the
bind-mount escape check directly on /dev/ptmx. The function responsible for
this is devpts_ptmx_path() which calls pts_path() which in turn calls
path_parent_directory(). Based on the above explanation,
path_parent_directory() will yield / as the parent directory for the
/dev/ptmx bind-mount and not the expected /dev. Thus, the kernel detects
that /dev/ptmx is escaping its bind-mount and will set /proc/<pid>/fd/<nr>
to /.

This patch changes the logic to first resolve any bind-mounts. After the
bind-mounts have been resolved (i.e. we have traced it back to the
associated devpts mount) devpts_ptmx_path() can be called. In order to
guarantee correct path generation for the slave file descriptor the kernel
now requires that a pts directory is found in the parent directory of the
ptmx bind-mount. This implies that when doing bind-mounts the ptmx
bind-mount and the devpts mount should have a common parent directory. A
valid example is:

mount -t devpts devpts /dev/pts
mount --bind /dev/pts/ptmx /dev/ptmx

an invalid example is:

mount -t devpts devpts /dev/pts
mount --bind /dev/pts/ptmx /ptmx

This allows us to support:
- calling open on ptmx devices located inside non-standard devpts mounts:
  mount -t devpts devpts /mnt
  master = open("/mnt/ptmx", ...);
  slave = ioctl(master, TIOCGPTPEER, ...);
- calling open on ptmx devices located outside the devpts mount with a
  common ancestor directory:
  mount -t devpts devpts /dev/pts
  mount --bind /dev/pts/ptmx /dev/ptmx
  master = open("/dev/ptmx", ...);
  slave = ioctl(master, TIOCGPTPEER, ...);

while failing on ptmx devices located outside the devpts mount without a
common ancestor directory:
  mount -t devpts devpts /dev/pts
  mount --bind /dev/pts/ptmx /ptmx
  master = open("/ptmx", ...);
  slave = ioctl(master, TIOCGPTPEER, ...);

in which case save path generation cannot be guaranteed.

CVE References

Christian Brauner (cbrauner) wrote :

Oh sorry, it would be great if we could cherry-pick these as this currently breaks users. Sorry for all the pre-release extra work this causes. :)


This bug is missing log files that will aid in diagnosing the problem. While running an Ubuntu kernel (not a mainline or third-party kernel) please enter the following command in a terminal window:

apport-collect 1755857

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

Changed in linux (Ubuntu):
status: New → Incomplete
Changed in linux (Ubuntu):
status: Incomplete → Confirmed
Joseph Salisbury (jsalisbury) wrote :

Are these patches only needed in Bionic, or do you need them in prior releases as well?

Changed in linux (Ubuntu Bionic):
importance: Undecided → High
status: Confirmed → Triaged
tags: added: bionic kernel-da-key
Christian Brauner (cbrauner) wrote :

TIOCGPTPEER was introduced in 4.13 so if we could backport it to the kernels we still support that would be great.
The patches will likely also be included in upstream stable updates.


Changed in linux (Ubuntu Bionic):
status: Triaged → Fix Committed
Launchpad Janitor (janitor) wrote :
Download full text (32.6 KiB)

This bug was fixed in the package linux - 4.15.0-13.14

linux (4.15.0-13.14) bionic; urgency=medium

  * linux: 4.15.0-13.14 -proposed tracker (LP: #1756408)

  * devpts: handle bind-mounts (LP: #1755857)
    - SAUCE: devpts: hoist out check for DEVPTS_SUPER_MAGIC
    - SAUCE: devpts: resolve devpts bind-mounts
    - SAUCE: devpts: comment devpts_mntget()
    - SAUCE: selftests: add devpts selftests

  * [bionic][arm64] d-i: add hisi_sas_v3_hw to scsi-modules (LP: #1756103)
    - d-i: add hisi_sas_v3_hw to scsi-modules

  * [Bionic][ARM64] enable ROCE and HNS3 driver support for hip08 SoC
    (LP: #1756097)
    - RDMA/hns: Refactor eq code for hip06
    - RDMA/hns: Add eq support of hip08
    - RDMA/hns: Add detailed comments for mb() call
    - RDMA/hns: Add rq inline data support for hip08 RoCE
    - RDMA/hns: Update the usage of sr_max and rr_max field
    - RDMA/hns: Set access flags of hip08 RoCE
    - RDMA/hns: Filter for zero length of sge in hip08 kernel mode
    - RDMA/hns: Fix QP state judgement before sending work requests
    - RDMA/hns: Assign dest_qp when deregistering mr
    - RDMA/hns: Fix endian problems around imm_data and rkey
    - RDMA/hns: Assign the correct value for tx_cqn
    - RDMA/hns: Create gsi qp in hip08
    - RDMA/hns: Add gsi qp support for modifying qp in hip08
    - RDMA/hns: Fill sq wqe context of ud type in hip08
    - RDMA/hns: Assign zero for pkey_index of wc in hip08
    - RDMA/hns: Update the verbs of polling for completion
    - RDMA/hns: Set the guid for hip08 RoCE device
    - net: hns3: Refactor of the reset interrupt handling logic
    - net: hns3: Add reset service task for handling reset requests
    - net: hns3: Refactors the requested reset & pending reset handling code
    - net: hns3: Add HNS3 VF IMP(Integrated Management Proc) cmd interface
    - net: hns3: Add mailbox support to VF driver
    - net: hns3: Add HNS3 VF HCL(Hardware Compatibility Layer) Support
    - net: hns3: Add HNS3 VF driver to kernel build framework
    - net: hns3: Unified HNS3 {VF|PF} Ethernet Driver for hip08 SoC
    - net: hns3: Add mailbox support to PF driver
    - net: hns3: Change PF to add ring-vect binding & resetQ to mailbox
    - net: hns3: Add mailbox interrupt handling to PF driver
    - net: hns3: add support to query tqps number
    - net: hns3: add support to modify tqps number
    - net: hns3: change the returned tqp number by ethtool -x
    - net: hns3: free the ring_data structrue when change tqps
    - net: hns3: get rss_size_max from configuration but not hardcode
    - net: hns3: add a mask initialization for mac_vlan table
    - net: hns3: add vlan offload config command
    - net: hns3: add ethtool related offload command
    - net: hns3: add handling vlan tag offload in bd
    - net: hns3: cleanup mac auto-negotiation state query
    - net: hns3: fix for getting auto-negotiation state in hclge_get_autoneg
    - net: hns3: add support for set_pauseparam
    - net: hns3: add support to update flow control settings after autoneg
    - net: hns3: add Asym Pause support to phy default features
    - net: hns3: add support for querying advertised pause frame by ethtool ethx
    - net:...

Changed in linux (Ubuntu Bionic):
status: Fix Committed → Fix Released
Brad Figg (brad-figg) wrote :

This bug is awaiting verification that the kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-bionic' to 'verification-done-bionic'. If the problem still exists, change the tag 'verification-needed-bionic' to 'verification-failed-bionic'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See for documentation how to enable and use -proposed. Thank you!

tags: added: verification-needed-bionic
Andy Whitcroft (apw) on 2019-02-14
tags: added: kernel-fixup-verification-needed-bionic
removed: verification-needed-bionic
Brad Figg (brad-figg) on 2019-02-14
tags: added: verification-needed-bionic
Andy Whitcroft (apw) wrote :

This bug was erroneously marked for verification in bionic; verification is not required and verification-needed-bionic is being removed.

tags: removed: verification-needed-bionic
tags: added: verification-done-bionic
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers